communication.serializer_plugins.pandas module

(De)serialization logic for pandas objects.

communication.serializer_plugins.pandas.pandas_deserialize_dataframe(obj, use_pickle, **_kwargs)

Function for deserializing pandas dataframe

Parameters:

• obj (bytes | dict[Hashable, Any]) – pandas dataframe to deserialize

• use_pickle (bool) – set to True to enable serialization fallback to pickle

• **_kwargs (Any) – optional extra keyword arguments

Raises:

ImportError – Object was serialized with parquet, but required dependencies for deserialization are missing.

Return type:

DataFrame

Returns:

deserialized dataframe

communication.serializer_plugins.pandas.pandas_deserialize_series(obj, **kwargs)

Function for deserializing pandas series

Parameters:

• obj (bytes | dict[Hashable, Any]) – pandas series to deserialize

• **kwargs (Any) – optional extra keyword arguments

Return type:

Series

Returns:

deserialized series

communication.serializer_plugins.pandas.pandas_serialize_dataframe(obj, use_pickle, **kwargs)

Function for serializing pandas dataframes

Attempt to use parquet for smaller serialized dataframe, but fallback to dictionaries otherwise.

Parameters:

• obj (DataFrame) – pandas object to serialize

• use_pickle (bool) – set to True to enable serialization fallback to pickle

• **kwargs (Any) – optional extra keyword arguments

Return type:

bytes | dict[Hashable, Any]

Returns:

serialized dataframe

communication.serializer_plugins.pandas.pandas_serialize_series(obj, **_kwargs)

Function for serializing pandas series

Parameters:

• obj – pandas series to serialize

• **_kwargs – optional extra keyword arguments

Returns:

serialized series

communication.serializer_plugins.pandas.register()

Register pandas serializer and deserializer.

Return type:

None